Predictors of Quality of Life in Treated Cushing's Disease
Predictors of Quality of Life in Treated Cushing's Disease
Percentages were used to describe pre-operative MRI findings, remission status, as well as comorbidities at baseline and follow-up. Mean and range were reported for age at diagnosis, pre-operative ACTH, pre-operative GC exposure and duration of follow-up; median and range were reported for pre-operative UFC and cortisol. Unpaired Students t-tests and chi-squared tests were used to compare outcomes based on remission status. Simple linear regression was used to identify clinical predictors of QoL in treated CD. Predictors included age at follow-up, sex, current BMI, comorbidities at follow-up, pre- and post-operative biochemical data and remission status. Multiple linear regression models A and B were constructed via forward stepwise selection to include variables of clinical and statistical significance on univariate analysis. Models were adjusted for (A) age at follow-up, sex, BMI, self-assessed recovery duration and post-TSS remission status and (B) age at follow-up, sex, hypopituitarism and post-TSS remission status (see Table S1A http://onlinelibrary.wiley.com/store/10.1111/cen.12521/asset/supinfo/cen12521-sup-0001-TableS1.docx?v=1&s=01018f82acb78da0b22014c10d493fe99ea672c1 for complete regression models). A second set of multiple linear regression models were constructed with self-assessed remission status as a variable (see Table S1B http://onlinelibrary.wiley.com/store/10.1111/cen.12521/asset/supinfo/cen12521-sup-0001-TableS1.docx?v=1&s=01018f82acb78da0b22014c10d493fe99ea672c1 for complete regression models). Model A was used to determine predictors of CushingQoL score, while model B was used to identify predictors of both HADS Depression and HADS Anxiety scores. Eight subjects were excluded from Model A on the basis of missing data (self-assessed recovery duration). The demographics and characteristics of these subjects were not dissimilar from that of the total cohort. Statistical significance was defined as a P value <0·05. All analyses were performed using stata version 11.2 (STATACORP, College Station, TX, USA) and Microsoft Excel (Microsoft Corporation, Redmond, WA, USA).
Statistical Analysis
Percentages were used to describe pre-operative MRI findings, remission status, as well as comorbidities at baseline and follow-up. Mean and range were reported for age at diagnosis, pre-operative ACTH, pre-operative GC exposure and duration of follow-up; median and range were reported for pre-operative UFC and cortisol. Unpaired Students t-tests and chi-squared tests were used to compare outcomes based on remission status. Simple linear regression was used to identify clinical predictors of QoL in treated CD. Predictors included age at follow-up, sex, current BMI, comorbidities at follow-up, pre- and post-operative biochemical data and remission status. Multiple linear regression models A and B were constructed via forward stepwise selection to include variables of clinical and statistical significance on univariate analysis. Models were adjusted for (A) age at follow-up, sex, BMI, self-assessed recovery duration and post-TSS remission status and (B) age at follow-up, sex, hypopituitarism and post-TSS remission status (see Table S1A http://onlinelibrary.wiley.com/store/10.1111/cen.12521/asset/supinfo/cen12521-sup-0001-TableS1.docx?v=1&s=01018f82acb78da0b22014c10d493fe99ea672c1 for complete regression models). A second set of multiple linear regression models were constructed with self-assessed remission status as a variable (see Table S1B http://onlinelibrary.wiley.com/store/10.1111/cen.12521/asset/supinfo/cen12521-sup-0001-TableS1.docx?v=1&s=01018f82acb78da0b22014c10d493fe99ea672c1 for complete regression models). Model A was used to determine predictors of CushingQoL score, while model B was used to identify predictors of both HADS Depression and HADS Anxiety scores. Eight subjects were excluded from Model A on the basis of missing data (self-assessed recovery duration). The demographics and characteristics of these subjects were not dissimilar from that of the total cohort. Statistical significance was defined as a P value <0·05. All analyses were performed using stata version 11.2 (STATACORP, College Station, TX, USA) and Microsoft Excel (Microsoft Corporation, Redmond, WA, USA).